Trying to ascertain the correct regular expression

Question

I am using Java and have a String, sum, whose content will be something like this:

A bunch of miscellaneous text
A bunch more miscellaneous text
Handicap Accessible *value*
More miscellaneous Text
Even more miscellaneous text

value can be Yes, No or None

I am trying to get the value of value with a regular expression. I can't just do a sum.replaceAll("^.*Handicap Accessible ","") because there are new lines and other characters which don't count in ".".

I am trying to use a regular expression but I can't get it right. Below is what I have tried, both with and without the backslashes. And notice that this is from java so I need to use two backslashes (\\):

    Pattern pat = Pattern.compile("Handicap Accessible \\([A-Za-z]*\\)");
    Matcher match = pat.matcher(sum);
    String newAccess = null;
    while (match.matches()) {
        newAccess = match.group(1);
        break;
    }

But when I print the value of newAccess it is always null. If I initialize newAccess to something else like "GLUB" then "GLUB" is what is printed at the end, meaning the matches loop is not being entered.

Any suggestions for the correct pattern to use?

yes (potentially, that is sometimes) to new lines. No to backslashes (Values are Yes, No, None). cricket_007 that would find the entire string but how would I extract it? And in the case when there is an error, the value, though it should not be, may be something else (again it shouldn't be but in an eror condition) — Tony, Jun 15 '16 at 01:55
The parenthesis in the @cricket_007 solution indicates a group so you can extract it from `matches.group(1)` The error values wouldn't be matched at all. Isn't that what you want? — Jorge Campos, Jun 15 '16 at 01:57
I will have an expected value: No, Yes or None, depending on the test and I have to verify that the value is as expected. I could say sum.contains("Handicap Accessible No") for instance. That would work — Tony, Jun 15 '16 at 01:59
Okay, so the regex you have in the question is a literal backslash, followed by a group of zero or more characters, then another backslash. That is why I asked. — OneCricketeer, Jun 15 '16 at 02:35
I always forget whether in a group (for a matching group) you use ( or \( ;-) — Tony, Jun 15 '16 at 12:31

Jorge Campos · Answer 1 · 2016-06-15T02:21:32.640

1

I would go for (took the input from @Tim's answer, thanks Tim):

String input = "A bunch of miscellaneous text\n" +
           "A bunch more miscellaneous text\n" +
           "Handicap Accessible None\n" +
           "Handicap Accessible Yes\n" +
           "More miscellaneous Text\n" +
           "Handicap Accessible No\n" +
           "Handicap Accessible somevalue\n" +
           "Even more miscellaneous text\n";
Pattern p = Pattern.compile("^Handicap Accessible (Yes|None|No)$", Pattern.MULTILINE);
Matcher m = p.matcher(input);
while ( m.find() ){
    System.out.println( "Value is: " + m.group(1) );
}

Following the @Qix suggestion I changed from Handicap Accessible (Yes|None|No) to "^Handicap Accessible (Yes|None|No)$", Pattern.MULTILINE it works better in a multiline text (instead of just \n)

I will output:

Value is: None
Value is: Yes
Value is: No

edited Jun 15 '16 at 02:21

answered Jun 15 '16 at 02:03

Jorge Campos

22,647
7
56
87

For completeness, pass `Pattern.MULTILINE` as a second parameter to `.compile()` and add `^` and `$` to the regex. This will enforce checking each line separately instead of a line such as `Handicap Accessible Yes Handicap Accessible No\n` having two results. – Qix - MONICA WAS MISTREATED Jun 15 '16 at 02:16
Eh? `Pattern.MULTILINE` will look at each line, so yes `^` and `$` will work just fine. [See for yourself.](https://ideone.com/xKmASr) – Qix - MONICA WAS MISTREATED Jun 15 '16 at 02:21
I deleted the comment because your previous comment said `m` as second parameter. Thanks for the suggestion – Jorge Campos Jun 15 '16 at 02:22
Ah yeah, apologies. I was thinking about Javascript's `RegExp`'s second constructor parameter. Had to double check ;) – Qix - MONICA WAS MISTREATED Jun 15 '16 at 02:22

score 0 · Accepted Answer · edited May 23 '17 at 12:15

0

Replace the match.matches() with match.find():

Pattern pat = Pattern.compile("Handicap Accessible ([A-Za-z]*)");
Matcher match = pat.matcher(sum);
String newAccess = null;
while (match.find()) {
    newAccess = match.group(1);
    break;
}

the difference between of them can be seen from Difference between matches() and find() in Java Regex. Meanwhile, "\\(" means your pattern need matches the character "(".

edited May 23 '17 at 12:15

Community

1
1

answered Jun 15 '16 at 01:59

Heng

96
1
5

Thanks. I know usually \ escapes a character, and I could never remember if you are supposed to use ( or \( for a group. Sounds like no \. I am going to try the suggestions above and let you guys know, – Tony Jun 15 '16 at 11:58
This worked when I used Pattern.MULTILINE and find ;-) – Tony Jun 15 '16 at 12:28

score 0 · Answer 3 · answered Jun 15 '16 at 02:07

Well, I think this is straightforward. Just apply the regex and print what is in the capture group:

    String input = "A bunch of miscellaneous text\n" +
                   "A bunch more miscellaneous text\n" +
                   "Handicap Accessible None\n" +
                   "More miscellaneous Text\n" +
                   "Even more miscellaneous text\n";

    String regex="Handicap Accessible\\s+(Yes|No|None)";
    Pattern pattern = Pattern.compile(regex);
    Matcher matcher = pattern.matcher(input);   
    String value=null;
    if (matcher.find()) {
        value=matcher.group(1);
    } else {
        throw new RuntimeException ("your regex is wrong dude!");
    }
    System.out.println(value);

This works too, but I also want it to print the bad value. Another thing I was toying with was WebElement ele = driver.findElement(By.xpath("//*[contains(text(), 'Handicap Accessible')]"); and then get the value of the text, and then it would be easier to manipulate Handicap Accessible, instead of getting the entire pageText(); — Tony, Jun 15 '16 at 12:30

Trying to ascertain the correct regular expression

3 Answers3